The present invention relates in general to the detection of trip-wires and in particular to the detection of trip-wires using optical techniques.
The reliable detection of trip-wires has been a problem that has plagued the military research community for some time. Military and law enforcement agencies currently do not have an effective way of detecting simple trip-wires. This problem is exacerbated during combat situations in which meticulous inspection of one""s pathway is often not possible. Civilian law enforcement groups, such as the Drug Enforcement Agency (DEA), have reported an increase in so-called xe2x80x9cbooby-trappedxe2x80x9d incidences involving their agents. It is often the case that when an illegal crop is identified, DEA personnel are placed at great risk during the secure phase of an operation in which booby-traps are searched out and disarmed.
A similar problem involves the detection and early warning of power-lines and hanging cables during certain helicopter missions. A particularly troublesome situation encountered by military pilots involves urban night missions in which the probability of a helicopter coming in contact with a power cable or wire is greatly increased.
The state of the art involving the detection of suspended wires and cables can be broken into two classifications, i.e., natural/ambient or active illumination schemes. Both methods are actually adaptations of a much broader application and consequently do not function very effectively. The natural illumination method (favored among most robotic scientists since it is the simplest to implement) involves applying various pattern recognition algorithms (PRAs) to transmitted video imagery from visible cameras mounted on a robot. The PRAs are designed to key in on and identify any xe2x80x9cfine-linexe2x80x9d structures that are present in the video scene. Unfortunately, this approach must overcome two fundamental problems. First, naturally illuminated 2-D video scenes do not convey the type of information necessary for PRAs to accurately distinguish between common straight edges (e.g., a sharp edge of a tabletop) and suspended wires. As a result, the false-alarm rate is often very high for all but the most simple of scenes. Second, by their very nature, trip-wires are designed to blend into their backgrounds and thus often do not exhibit the necessary contrast needed for PRAs to xe2x80x9ckey onxe2x80x9d.
Active illumination methods (sometimes termed 3-D laser imaging/Doppler) use a pulsed laser to illuminate an extended target by optically scanning out a 2-D area using a particular scan. Coincident sensors are then used to record the position and time delay in the scattered signal. A pseudo image is generated that gives rough dimensions and distance to the illuminated object. Such systems are currently deployed and undergoing evaluation on various military helicopters. These systems are both extremely complex and expensive to deploy. Furthermore, this technique is only effective in identifying targets that have reasonably large extended areas and are inefficient in identifying objects that possess small geometric cross-sections, such as hanging wires or cables.
Both active and passive techniques as outlined above are deficient in their approaches because they cue on features that are not entirely unique to the target of interest, i.e., trip-wires and/or cables.